Article Complete
AIM
Huawei’s AI Infrastructure Play is Hard to Ignore at MWC 2026
Huawei’s AI Infrastructure Play is Hard to Ignore at MWC 2026

Image by Smruti Biswal

AI-nt to Ignore

Huawei’s AI Infrastructure Play is Hard to Ignore at MWC 2026

Huawei provides AI infrastructure, industry models, agent platforms, and applications to address sector-specific challenges.

Published Updated

Key takeaways

  • Huawei positions itself as a leader in AI infrastructure at MWC 2026.
  • Emphasize the need for sovereign AI systems in telecom, reducing dependency on global vendors.
  • Introduce advanced optical networks to enhance AI capabilities and network performance.
  • Highlight the importance of high bandwidth and low latency for future telecom applications.
5 min

With a surge in AI applications, telcos are now repositioning themselves as providers of AI infrastructure. Sovereignty and control over data, models, supply chains, infrastructure, and networks have become the new currency. 

This has been the running theme at the ongoing Mobile World Congress (MWC) 2026 in Barcelona, where the conversations are focused largely on AI infrastructure over handsets.

Huawei used MWC 2026 to argue that AI cannot scale on legacy networks. The company has unveiled an AI-centric all-optical target network, expanded its cloud strategy, and released an industrial networks report. 

Together, these moves place Huawei at the centre of a growing debate about whether telecom operators should build sovereign AI systems without relying on global infrastructure vendors.

Huawei’s pitch focused on rebuilding the backbone of the telecom industry. Speaking at a product launch, Bob Chen, President of Huawei’s Optical Business Product Line, stated, “Huawei advances next generation optical network solutions in two directions: AI for Networks and Networks for AI.”

The AI for Networks approach is used to power smart fibre sensing, boost network capability and customer experience, streamline operations and maintenance, and cut energy use. 

Meanwhile, the Networks for AI strategy focuses on upgrading network performance. In effect, Huawei is using AI to optimise networks while upgrading those networks to support AI workloads.

Huawei tied this shift to higher bandwidth and lower latency, introducing 400G and 800G (gigabits per second) backbone upgrades and proposing latency tiers of 1 ms, 3 ms, and 5 ms for metro, regional, and national networks. These targets align with AI inference demands at the edge.

At the Green All-Optical Network Forum, Kim Jin, Vice President of Huawei Optical Business Product Line, said that the AI era brings operators a once-in-a-decade opportunity. “This will keep improving user experience, unlock more network value, and achieve win-win growth in the AI era.”

A Step Towards Physical AI

Alongside discussions on network and cloud infrastructure, developments in robotics and embodied AI also took centre stage at MWC 2026. HONOR, formerly a Huawei sub-brand and independent since 2020, advanced its AI vision with a suite of connected devices that extend intelligence from software and networks into physical form. 

At the event, the company offered a preview of its Robot Phone, which blends motion control with AI interaction. 

Built with a robot-grade motion system, the Robot Phone combines “embodied AI interaction with robot-grade motion and cinematic imaging capabilities,” according to the company’s statement. The device supports all-angle AI video calling, responds with expressive gestures such as nods and head shakes, and can even “dance to the beat of music” as part of its interaction model. 

HONOR also showcased its first humanoid robot at the congress, marking the company’s first entry into consumer-oriented physical robotics. It said these robots will initially focus on practical scenarios such as shopping assistance, workplace inspection and supportive companionship, signalling an ambition to integrate embodied AI into everyday environments. 

James Li, CEO of HONOR, framed this direction as part of a broader strategy to integrate AI across hardware and software. “With human-centric as our lighthouse, we navigate the growth of AI through the two beams of IQ and EQ,” he said, underscoring the company’s push to unify intelligent cognition with expressive, physical interactions. 

Rebuilding the AI Backbone

Huawei extended its infrastructure theme into cloud and enterprise. At the Huawei Cloud Summit, the company unveiled its Industry AI Foundry and the next-generation hybrid cloud offering, Huawei Cloud Foundation. It also introduced CodeArts, an AI-powered coding agent designed to integrate code models, development tools, and automation.

Peter Zhou, CEO of Huawei Cloud, said cloud and AI sit at the core of Huawei’s strategy and drive steady investment in research and development. He stated that Huawei Cloud remains committed to delivering secure and reliable services while collaborating to cultivate an open ecosystem.

Huawei framed cloud as the foundation for industry AI adoption. The company said it provides AI infrastructure, industry models, agent platforms and applications to address sector-specific challenges.

In manufacturing, the company added that effective AI implementation requires strong foundation models and high-quality industry data. It positioned itself as a partner to governments and operators seeking to embed AI into national infrastructure.

Telcos Push Back with Sovereign AI Plans

Other operators at MWC 2026 echoed the theme of sovereignty. 

South Korean wireless telecom operator SK Telecom announced its ‘Sovereign AI Package’ and plans to build 1GW-class hyperscale AI data centres across the country. CEO Jung Jaihun said, “Competitiveness in the AI era depends not only on technology, but also on how infrastructure is designed and interconnected. We will work together with global telecom operators to build a trusted AI infrastructure and service ecosystem.”

The company also plans to upgrade its sovereign foundation model to over 1 trillion parameters and expand AI-RAN deployment. The operator aims to control both computing and models within national boundaries.

SoftBank introduced its Telco AI Cloud vision. The company said it will integrate GPU cloud, AI-RAN-based edge platforms, and software for AI data centres. SoftBank stated that it seeks to evolve beyond a telecommunications operator and become an AI infrastructure provider.

Semiconductor suppliers also signalled their role in the shift. Swedish company Sivers Semiconductors showcased SATCOM beamforming chips and flat-panel arrays for high-speed wireless and satellite connectivity. These components support low-latency backhaul and distributed AI services.

Telecom operators now clearly want to move up the value chain. The question remains, who becomes indispensable to their operation?

Membership

Stay ahead with AI insights.

Unlimited access to premium articles, exclusive analysis, and the stories shaping India's AI economy.

Subscribe now

AIM Americas

1190 Miraloma Way Suite P, Sunnyvale, CA 94085

AIM India

1st Floor, Sakti Statesman, Bellandur, Bengaluru – 560103

© 2026 AIM. All rights reserved.